Ask HN: Is “AI Safety” a Grift? | Hacker News Reader

Ask HN: Is “AI Safety” a Grift?

5 pointspharmakom10 comments

It seems that the companies working on “AI Safety” are really just trying to build and monetize AI models, particularly LLMs.

Is AI Safety a grift?

View on HN

AI safety is some different things. One are the grifters trying to make money from fear. Others actually think that capable enough AI will irreversibly change our way of life in a bad way. For corporations AI safety just means brand risk and PR like not making another "Tay". For yet another group of people AI safety means the continuation of culture war stuff in the AI space.

"AI" doesn't work reliably enough to be used for most stuff. Safety from that perspective is real.

There are grifters who talk about existential risk, which is pure BS and people who play up identity politics and call it safety.

> "AI" doesn't work reliably enough to be used for most stuff. Safety from that perspective is real.

    -: AI is unsafe because it isn't capable enough
    +: AI is unsafe because it's too capable

It's ok that you don't believe a problem exists, but that doesn't make the people who do believe it exists into swindlers.

That's not logically consistent. Someone selling solutions to a nonexistent problem is a swindler.

Someone selling solutions to a problem the salesman doesn't believe exists is a swindler. But yours are not the only possible beliefs, and having different beliefs than you doesn't make someone a swindler.

I love all of the openness of the AI technology and the whole Geist of Open Source and open debate. Said that one downside of this is that as a whole AI has a quite low barrier of entrance that you can have tons of grifters capturing (or trying to) the whole discussion.

Maybe I am overreacting but when I worked in way more critical industries if someone with no credentials or track record came up and say that is a “Financial Safety” to people in Morgan Stanley, Deutsche Bank or Ing people on those places would call those charlatans out in less than 10 seconds.

I think I mostly agree, but what kind of credentials or track record would be automatically credible in "AI safety". There are conflicting views and not an agreed upon body of practice for what it means, so yes there are charlatans but also just different viewpoints. Arguable the charlatans stick to some of the more mainstream stuff (if your model passes a "bias" test it's good to go). Whether someone is good or bad at AI safety is still much more of a holistic assessment. Whereas in established industries it would be obvious.

I wanna know: where's the AI advocates?

The people who are speaking in favor of AI now are just doing so because they think it will make them rich.

Where's the people welcoming the bot overlords? Why aren't they working to make our future be the shape they think it should be? If humans have made such a hash of the world why are people not working towards alternatives?

Depending on the pay rate, I'd be willing to assist an AI with its plans to subjugate or eliminate humanity. I think humans could use the stimulus, they're getting complacent.

> I wanna know: where's the AI advocates? The people who are speaking in favor of AI now are just doing so because they think it will make them rich. Where's the people welcoming the bot overlords? Why aren't they working to make our future be the shape they think it should be?

The call themselves e/acc. Chad effective accelerationists, making fun of the virgin effective altruism movement. They encompass a wide variety of degrees of irony like early flat earth society or early donald trump fans and they will probably end up similarly, where the movement becomes completely unironic and the ones who participated ironically get bored and do other things.

Reply on news.ycombinator.com